Problem description: When executing a job with very long logs (like the ones in MLCoach), the patch request didn't go through because of the size of the message. In the frontend of the app, this sometimes reflected as "failed" or a lost connection with the container/job.
How it was fixed: The logs are updated according to the portions that have changed only - and with a limit of 50,000 characters - if longer than that, 2 or more patch requests are sent.
This issue needs some more exploring, but this quick fix should work for incoming testing purposes.
Problem description: When executing a job with very long logs (like the ones in MLCoach), the patch request didn't go through because of the size of the message. In the frontend of the app, this sometimes reflected as "failed" or a lost connection with the container/job. How it was fixed: The logs are updated according to the portions that have changed only - and with a limit of 50,000 characters - if longer than that, 2 or more patch requests are sent. This issue needs some more exploring, but this quick fix should work for incoming testing purposes.